How to manually clean up a failed replication task if the vRanger host has failed or rebooted dur (4042676)

There may be times where a replication task has not finished executing all the subtasks due to a failure or a reboot of the vRanger host.

In cases like this it may be necessary to manually roll back any changes that were made prior to the network failure before resuming replication.

As data is sent to the destination host (via the VA), blocks in the destination disk are written to the undo file on the VA before they are overwritten by the changed data. If replication fails and an undo becomes necessary, then the original destination disk blocks are read from the undo file and written to the destination disk to rollback the failed replication. This is a key function designed to provide resiliency in the face of a network failure; if there is a network failure during the replication pass, the destination VM is not corrupted by incomplete data.

It may be necessary to manually roll back any changes that were made prior to the network failure before resuming replication.

After network connectivity has been restored, perform the following:

For VA Network Replications:

1. ssh into the target VA.
2. Navigate to the directory that contains the vz* files for the VM that is being replicated. Typically the directory is located /scratch/sdb/<UUID>. If multiple VM's are using the VA for replications you will need to find the specific folder that is storing the VM's vzundo files. If the vzundo file does not exist, skip to step 5.
3. Once in the directory, run 'chmod 777 vzundo-script'
4. Execute the script by running './vzundo-script'
5. In your VI Client, revert and remove the "vRanger" snapshot on the Replica VM.
6. Re-run your replication task.
    

For VA HotAdd Replications:

1. Unmount the VM's disk from the VA. This can be accomplished through the VI Client. Remove the HotAdded VM's Hard Drive from the VA's settings. Only Remove the VMDK from the VA. Do NOT delete the files from disk.
2. ssh into the target VA.
3. Navigate to the directory that contains the vz* files for the VM that is being replicated. Typically the directory is located /scratch/sdb/<UUID>. If multiple VM's are using the VA for replication you will need to find the specific folder that is storing the VM's vzundo files. If the vzundo file does not exist, skip to step 6.
4. Once in the directory, run 'chmod 777 vzundo-script'
5. Execute the script by running './vzundo-script'
6. In your VI Client, revert and remove the "vRanger" snapshot on the Replica VM.
7. Re-run your replication task.